Schneider-Thoma Johannes, Hamza Tasnim, Chalkou Konstantina, Siafis Spyridon, Dong Shimeng, Bighelli Irene, Hansen Wulf-Peter, Scheuring Elfriede, Davis John M, Priller Josef, Baumann Pierre, Conley Robert, Cordes Joachim, Kelly Deanna, Kluge Michael, Kumra Sanjiv, Lewis Shôn, Meltzer Herbert Y, Naber Dieter, Schooler Nina, Volavka Jan, Wahlbeck Kristian, Salanti Georgia, Leucht Stefan
Department of Psychiatry and Psychotherapy, School of Medicine and Health, Technical University of Munich, Munich, Germany; German Center for Mental Health, Munich, Germany.
Institute of Social and Preventive Medicine, University of Bern, Bern, Switzerland; Graduate School for Health Sciences, University of Bern, Bern, Switzerland.
Lancet Psychiatry. 2025 Apr;12(4):254-265. doi: 10.1016/S2215-0366(25)00001-X. Epub 2025 Feb 26.
Clozapine is recommended by national and international guidelines for people with treatment-resistant schizophrenia. However, available meta-analyses of randomised controlled trials have not shown superior efficacy of clozapine when compared with other second-generation antipsychotics, with heterogeneity identified between the original studies. We aimed to use individual patient data (IPD) to account for potential reasons of variability and to synthesise an adjusted estimate for the difference in efficacy between clozapine and other second-generation antipsychotics for treatment-resistant schizophrenia.
In this systematic review and IPD meta-analysis, we searched the Cochrane Schizophrenia Group's Study-Based Register from inception to Jan 24, 2024, and previous reviews for blinded randomised controlled trials comparing clozapine with other second-generation antipsychotics in participants with treatment-resistant schizophrenia. Trials were eligible if they included patients with treatment-resistant schizophrenia and had a duration of at least 6 weeks. IPD were requested from trial investigators. The primary outcome was change in overall schizophrenia symptoms as measured by the Positive and Negative Syndrome Scale (PANSS) between clozapine and other second-generation antipsychotics after 6-8 weeks of treatment. The effect size measure for the primary outcome was mean difference with 95% credible interval (CrI). We fitted a Bayesian random-effects IPD meta-regression model that included duration of illness, baseline severity, and sex as potential prognostic factors or treatment effect modifiers. Confidence in the evidence was assessed using Grading of Recommendations, Assessment, Development, and Evaluation (GRADE). People with lived experience of mental illness were involved in this study. This study is registered with PROSPERO, CRD42021254986.
We screened 13 876 references and included 19 studies with data for 1599 participants; IPD were available for 12 of 19 trials (n=1052; mean age 37·67 years [SD 11·24; range 10-66]; 348 [33·08%] women and 704 [66·92%] men). Data on ethnicity were not available. The estimated mean difference in change from baseline PANSS total score was -0·64 points (95% CrI -3·97 to 2·63; τ=2·68) in favour of other second-generation antipsychotics. Shorter duration of illness and higher baseline severity were prognostic factors associated with a larger reduction in symptoms, but neither those factors nor sex were found to modify the relative effect between clozapine and other second-generation antipsychotics. The confidence in the evidence was graded as very low.
This IPD meta-analysis found a small and uncertain advantage of other second-generation antipsychotics, mainly olanzapine and risperidone, and so did not provide evidence for superior efficacy of clozapine compared with other second-generation antipsychotics in treatment-resistant schizophrenia. It is limited by unavailability of IPD for some studies, uncaptured sources of variance, and uncertainty due to premature study discontinuation. Given the side-effects of clozapine, the observed uncertainty regarding clozapine's superiority warrants prudent use and further research.
German Ministry of Education and Research.
氯氮平被国内外指南推荐用于治疗难治性精神分裂症患者。然而，现有的随机对照试验的荟萃分析并未显示氯氮平与其他第二代抗精神病药物相比具有更高的疗效，且原始研究之间存在异质性。我们旨在使用个体患者数据（IPD）来解释变异性的潜在原因，并综合得出氯氮平与其他第二代抗精神病药物治疗难治性精神分裂症疗效差异的校正估计值。
在这项系统评价和IPD荟萃分析中，我们检索了Cochrane精神分裂症研究组基于研究的注册库，从其建立至2024年1月24日，并检索了之前的综述，以查找比较氯氮平与其他第二代抗精神病药物治疗难治性精神分裂症患者的盲法随机对照试验。如果试验纳入了难治性精神分裂症患者且持续时间至少为6周，则该试验符合纳入标准。我们向试验研究者索取IPD。主要结局是治疗6 - 8周后，氯氮平与其他第二代抗精神病药物相比，通过阳性和阴性症状量表（PANSS）测量的精神分裂症总体症状的变化。主要结局的效应量指标是均值差及95%可信区间（CrI）。我们拟合了一个贝叶斯随机效应IPD荟萃回归模型，该模型将病程、基线严重程度和性别作为潜在的预后因素或治疗效果修饰因素。使用推荐分级、评估、制定和评价（GRADE）对证据的可信度进行评估。有精神疾病生活经历的人参与了本研究。本研究已在国际前瞻性系统评价注册库（PROSPERO）注册，注册号为CRD42021254986。
我们筛选了13876篇参考文献，纳入了19项研究，共1599名参与者的数据；19项试验中的12项（n = 1052；平均年龄37.67岁[标准差11.24；范围10 - 66]；348名[33.08%]女性和704名[66.92%]男性）可获得IPD。没有种族数据。从基线PANSS总分变化的估计均值差为 - 0.64分（95% CrI - 3.97至2.63；τ = 2.68），有利于其他第二代抗精神病药物。病程较短和基线严重程度较高是与症状减轻幅度较大相关的预后因素，但未发现这些因素以及性别会改变氯氮平与其他第二代抗精神病药物之间的相对疗效。证据的可信度被评为非常低。
这项IPD荟萃分析发现其他第二代抗精神病药物（主要是奥氮平和利培酮）有微小且不确定的优势，因此没有提供证据表明氯氮平在治疗难治性精神分裂症方面比其他第二代抗精神病药物具有更高的疗效。它受到一些研究无法获得IPD、未捕捉到的变异来源以及因研究过早终止导致的不确定性的限制。鉴于氯氮平的副作用，观察到的关于氯氮平优越性的不确定性值得谨慎使用并进一步研究。
德国教育和研究部。
